Bilingual competence and bilingual proficiency in child development

Preface ix Acknowledgments xiii Abbreviations xv 1 Introduction: The Problem of Language Acquisition When There Are Two 1 1.1 Bilingual Profi ciency and Bilingual Competence 3 1.2

Child Development

All rights reserved No part of this book may be reproduced in any form by any electronic or mechanical means (including photocopying, recording, or information storage and retrieval) without permission in writing from the publisher

For information about special quantity discounts, please email special_sales@mitpress.mit.edu This book was set in Sabon by Toppan Best-set Premedia Limited Printed and bound in the United States of America

Library of Congress Cataloging-in-Publication Data

Francis, Norbert

Bilingual competence and bilingual profi ciency in child development / Norbert Francis

p cm

Includes bibliographical references and index

ISBN 978-0-262-01639-1 (alk paper)

1 Bilingualism in children 2 Language acquisition 3 Competence and performance (Linguistics)

Preface ix

Acknowledgments xiii

Abbreviations xv

1 Introduction: The Problem of Language Acquisition When There Are Two 1

1.1 Bilingual Profi ciency and Bilingual Competence 3

1.2 Knowledge That Outstrips Experience 10

1.3 Modularity 11

1.4 A Study of Indigenous-Language Bilingualism in Mexico 16

1.5 Looking Ahead: Overview of the Chapters 20

2 Bilingualism in School 25

2.1 When Second Language Learning Is Not Optional 27

2.2 Bilingualism, Diglossia, and Literacy 29

2.3 A Componential Approach to Language Ability Solves a Practical Problem in Second Language Learning 33

2.4 New Democracy in South Africa: The Challenge of a Multilingual Language Policy 35

2.5 A Possible Counterexample from North Africa 38

2.6 Program Design Based on a Concept from Sociolinguistics 44

3 The Debate on the Nature of Bilingual Profi ciency: Distinguishing between

Different Kinds of Language Ability 49

3.1 First Language and Second Language in Literacy Learning 51

3.2 Concepts of Bilingual Profi ciency: Background to the Debate 53 3.3 A Proposed Modifi cation of Cummins ’ s Model 56

3.4 Literacy Learning at the San Isidro Bilingual School: A Follow-Up Study 61

3.5 Comparing Results from Both Languages 63

3.6 Using the New Model to Describe Different Kinds of

Interdependence 68

3.7 Components and Connections 76

4 Componential Approaches to the Study of Language Profi ciency 79

4.1 Vygotsky and Luria: The Concept of “ Inner Speech ” 81

4.2 Metacognition: Language at the Service of Higher-Order Thinking 85 4.3 Compartmentalization of the Bilingual Mind 88

4.4 Bilingualism as a Showcase for the Internal Diversity of Language Profi ciency 94

4.5 Advancing the Research Program on Bilingualism: The Need for Clarity and Refl ection 101

5 Research on the Components of Bilingual Profi ciency 107

5.1 Maximum Imbalance in Bilingualism 109

5.2 Separation of the Linguistic Subsystems 115

5.3 How Bilingual Speech Constitutes Evidence of Language

Separation 118

5.4 Contradictions of an Integrativist Approach 124

5.5 A Bilingual Version of the Tripartite Parallel Architecture 125

5.6 More Opportunities for Research on Uneven Development 132

6 The Critical Period, Access to Universal Grammar in First and Second Language, and Language Attrition 141

6.1 Overview of the Chapter 142

6.2 The Concept of Language Attrition 144

6.3 What the Research Says about First Language Attrition 146

6.4 The Critical Period Hypothesis 151

6.5 Is Second Language Competence Universal Grammar –

Constrained? 159

6.6 Acquisition and Learning in the Second Language 166

6.7 A Wider Discussion: Applying Concepts to New Research 171

7 An Analysis of Academic Language Profi ciency 177

7.1 Secondary Discourse Ability + Metalinguistic Awareness 179

7.2 The Development of Narrativization and Levels of Narrative

Ability 183

7.3 Language Development — Grammar 187

7.4 Access to Shared Academic Profi ciencies in Biliteracy 192

7.5 Linking Secondary Discourse Ability and Metalinguistic Awareness at the Discourse, Sentence, and Word Levels 197

8 Metalinguistic Awareness, Bilingualism, and Writing 203

8.1 Metalinguistic Development and Bilingualism 204

8.2 Metalinguistic Awareness in Literacy and Second Language Learning 206 8.3 A Study of Children ’ s Perceptions of Focus on Form 207

8.4 Children ’ s Development of a Refl ective Posture toward Writing: Results from Spanish 210

8.5 Metalinguistic Awareness as a Component of Literacy Ability — Writing

in Particular 213

8.6 Possible Implications for Teaching Writing Skills 217

8.7 Children ’ s Development of a Refl ective Posture toward Writing: Results from Nahuatl 218

8.8 The Revision/Correction Assessment in Nahuatl 220

8.9 A Comparison of Performance between the Languages 222

8.10 Internal Resources and External Factors 223

8.11 Applying Different Kinds of Knowledge in Literacy Development 227

9 Metalinguistic Awareness, Bilingualism, and Reading 231

9.1 Modular Approaches to the Study of Reading 232

9.2 A Study of Focus on Form in Reading 234

9.3 The Development of a Refl ective Posture toward Reading

Comprehension 237

9.4 One Way in Which Children Learn to Use Context Strategically 240 9.5 Future Research on Literacy Learning, Metalinguistic Awareness, and Bilingualism 244

9.6 Does the Use of Context Contradict Modularity in Reading? 247

10 Conclusion: Results and Prospects 253

10.1 Parts to Whole: What ’ s Natural and What ’ s Unnatural in Language Learning? 254

10.2 Versions of Modularity and Pending Questions in Bilingual Research 266 10.3 Language Diversity, Cognition, and Culture 275

Appendix 1 Assessment of Metalinguistic Awareness Related to

Bilingualism 281

Appendix 2 Indices of Additive Bilingualism 289

Appendix 3 Early Childhood Borrowing and Codeswitching 295

Appendix 4 Writing Samples, including the Assessment of Revision/

This book is about the development of bilingual profi ciency and the different kinds

of underlying competence that come together in making up its component parts When two or more languages are part of a child ’ s world, we have a rich opportunity

to learn something about language in general and about how the mind works The same is true (some opportunities richer, others less so) for bilingualism in adults This explains in part why recent years have seen such an upsurge of interest in this area of research We will barely lift the cover on this voluminous body of investiga-tion In fact, we will restrict ourselves mainly to problems of language ability (pro-

fi ciency) when children use two languages for tasks related to schooling, especially

in learning how to read and write

Describing kinds of knowledge (competence) as “ underlying ” involves no idea that there is anything deep or occult about them Rather, it seems like a good way

to begin to frame some of the problems of language use — thinking about, for example, what the component parts of bilingual profi ciency might be so as to understand it better

Our main concern in looking at the research discussed here will be the questions that second language and bilingual educators ask This includes research that addresses issues of competence: how it develops, how knowledge is organized men-tally, and how it is processed As we get a better idea about the knowledge and processing components that come together in performance, we should better under-stand how two languages are used for different purposes At the same time, fi ndings from research specifi cally oriented toward aspects of learning and teaching pose interesting problems for other applied subfi elds, and even for work on theoretical models

That language learning and literacy might be enriched by including second guages, alongside the use of fi rst languages, is one idea that we will explore Another research proposal that makes this idea somewhat more interesting (because as it stands it ’ s rather unremarkable) is that this manner of language inclusion might apply without exception, for example, in school But not all languages (i.e., the

lan-people who speak them) have access to the same resources for developing learning materials and texts, and this clearly imposes certain limitations and practical con-straints So the question might better be formulated like this: how and to what extent might an inclusive bilingual or multilingual educational approach be applied even in cases where distribution of resources is sharply unequal?

The discussions about child bilingualism are all based on research or take research-based proposals as a beginning framework They also take a direction that

is infl uenced by a point of view, or just a view, that in some ways is more like a long view of where things might lead us somewhere down the road, someday For now, it would be safe to say that most researchers in the fi eld of bilingualism today have concluded that exclusionary (deliberately monolingual) approaches to lan-guage learning result in one or another kind of missed opportunity: one kind for children who already know the customary or offi cial language of instruction when they enter elementary school, and a related but different kind for children who know

a language that isn ’ t used for instruction and who need to learn a second One perspective on this state of affairs takes very seriously the following possibility: that scientifi c studies of language learning might contribute to clearing away unnecessary limitations on human development related to knowledge of one language or another

In theory, we would like to be able to say, it shouldn ’ t matter what language a child knows or knew fi rst

Much of the research reported on here comes from a project on child ism in which an indigenous language is part of the picture The full picture, when

bilingual-an indigenous or minority vernacular comes into contact with a national or offi cial language, more often than not involves missed opportunities of signifi cant propor-tion A major UNICEF-sponsored study of the developmental potential for children

in developing countries (Grantham-McGregor et al 2007) reminds us of this ing asymmetry in the world today The researchers estimated, conservatively, that over 200 million children in these countries are affected by loss of cognitive poten-tial In early childhood, contributing factors included extreme poverty, poor health, and poor nutrition, along one dimension, and defi cient care and impoverished experience along another Together, according to the authors, these factors predict not only low attendance and achievement in school, but overall attenuated cognitive attainment on a vast scale With 50% of the world ’ s population sharing about 1%

endur-of global wealth (Davies et al 2006), the 200 million estimate may correspond only

to the infant population at greatest risk of lost learning potential

Concentrated in the most marginalized and impoverished regions of sub-Saharan Africa, Latin America, and southern Asia, these children often are speakers of indigenous and minority languages with limited access to their respective national/offi cial languages, and even less so to instruction in a language they understand Despite signifi cant advances since the year 2000, particularly in Latin America, girls

are much more likely to suffer from limited access to a language of schooling, worldwide, than boys (UNESCO 2007) These unhappy correlations are not the subject of this book But, as certain related and tangential issues are examined in the coming chapters, we might want to consider adding this language-learning cir-cumstance to the ones cataloged by the UNICEF investigators as a predictor of academic failure during the middle childhood years A number of different kinds of extreme social and economic disparity have been shown to be serious obstacles to school achievement; and the unequal distribution of language-learning resources is one among them that needs to be better understood In the end, all of this should prompt us to think about the language policy implications that research on bilin-gualism and second language learning has called attention to over the years

Many people are owed thanks for helping shape the research project that I will report on here, fi rst and foremost the bilingual elementary students with whom I had the privilege of working in a number of rural communities in Central Mexico

It is to them and their families that the fi rst acknowledgment is most sincerely extended As a guest and visitor to their communities, I want to express my gratitude for the hospitality and generous understanding that they extended to me, always Thanks go especially to the teachers and principals of the following schools for access to their classrooms and for the many interesting discussions about bilingual-ism and language learning that we shared: the Escuela Xicoht é ncatl (municipality

of San Pablo del Monte, Tlaxcala state), the elementary schools of Pozuelos and Santa Teresa (municipality of Cardonal, Hidalgo state), and the elementary schools

of San Isidro and Uringuitiro (municipality of Los Reyes, Michoac á n state)

I would like to thank the colleagues and coworkers who have collaborated as coauthors on a number of reports published over the years, and who have been especially close to the work of the project: Rainer Enrique Hamel, Pablo Rogelio Navarrete G ó mez, Rafael Nieto Andrade, Jon Reyhner, and Phyllis Ryan Especially helpful as well were valuable consultations and extended discussions with Antonio Carrillo Avelar, Pedro Aztatzi Rugerio, Colin Baker, Rebeca Barriga Villanueva, Hintat Cheung, Karen Dakin, Yi Jhen Du, Kerim Friedman, Chieh-Fang Hu, Kent Johnson, Sam-Po Law, John McClure, Mercedes Montes de Oca Vega, Ishmael Munene, Timothy Murphy, Carol Myers-Scotton, Judith Oller Badenas, Akiyo Pahalaan, Kent Parks, Charles Perfetti, Sungok Serena Shim, Navin Kumar Singh, Tasaku Tsunoda, Chih-Hsiung Tu, Monkol Tungmala, Jiang Xia, Jin Xue, Jia-Ling Yau, and Emiko Yukawa I owe special gratitude to Juan Carlos Sierralta and the other staff members of the Direcci ó n General de Educaci ó n Ind í gena (DGEI), Mexico ’ s national department of indigenous education; my thesis director when I was a student at the UNAM, Martha Corenstein Zaslav; and, in Tlaxcala and neighboring Puebla in particular, Mar í a de Carmen Flores V á zquez, Enriqueta Vicenta Saucedo, Filiberto P é rez, Angel P é rez, Floriberto P é rez, Reyes Arce, Mar í a

Fernanda Magdalena Arce, V í ctor Arce Luna, Rub é n Sanchez, Avelino Zepeda, Miguel Zepeda, Trinidad Zepeda, and Scott Hadley, in addition to other members and friends of the Seminario de Estudios Modernos y de Cultura Acallan (SEMYCA)

in San Miguel Canoa As is customary and required, all persons mentioned from whom a language sample was taken are indicated with a pseudonym, for which I ask their forgiveness Thanks also go to Natalia and Zoraida for their patience and forbearance during the years of research and writing

Portions of earlier versions of chapters appeared in the following journals:

Applied Linguistics , International Journal of Applied Linguistics , International Journal of Bilingual Education and Bilingualism , International Journal of Bilingual- ism , Language Awareness , Language Learning , Language Problems and Language Planning (where “ Democratic Language Policy for Multilingual School Systems ” appeared in vol 19, pp 211 – 230), and Linguistics and Education

I extend great appreciation to the anonymous reviewers of this book for their observations, suggestions, and corrections, and for posing hard questions and chal-lenging critiques I am also grateful to MIT Press editor Ada Brunstein and her assistant Marc Lowenthal for keeping everything together for me, in more ways than one, and to editors Sandra Minkkinen and Anne Mark who did the same for this book

Finally, thanks go to the US/Mexico Fund for Culture, the Ford Foundation, and the Offi ce of Grants and Contracts of Northern Arizona University for supporting

an earlier phase of the work that started things going, many years back

ASL American Sign Language

BICS Basic Interpersonal Communication Skills

CALP Cognitive Academic Language Profi ciency

CLI Cross-Linguistic Interface

CPS Central Processing System

CS Conceptual Structure

CS ↔ L a+b Conceptual Structure ↔ linguistic structures interface

CUP Common Underlying Profi ciency

SLI Specifi c Language Impairment

SS Syntactic Structure

TE translation equivalent

ToM Theory of Mind

TPA Tripartite Parallel Architecture

Introduction

The Problem of Language Acquisition When There Are Two

Explaining how language acquisition unfolds in young children continues to elude consensus among investigators, even when we consider only one mother tongue Children exposed to two languages, beyond a minimum threshold, develop bilingual competence and bilingual profi ciency to some degree, usually to a degree that is surprising to both the casual observer and the student of language development Contact with more than two languages may result in multilingual competence and multilingual profi ciency Increasingly considered as a normal and even typical devel-opmental outcome, child bilingualism is attracting growing interest among investi-gators from many fi elds Part of the interest stems from the marvel and admiration that adults, including researchers, experience when we listen to young children alternate fl uently between one language and another

Major lines of research have opened up around several important questions:

• How might second language (L2) learning differ from fi rst language (L1) tion? Implied here is a kind of “ sequential ” bilingualism Is L2 learning different from monolingual L1 acquisition if it begins in early childhood, in middle child-hood, or in the post – elementary school years?

• If bilingualism is “ simultaneous ” (i.e., if two fi rst languages develop during early childhood), how might it differ from monolingual development? And how might it differ from “ sequential ” bilingualism?

• It is commonly observed that L2 grammatical development in older learners is uneven while L1 development is uniform Might this difference apply as well to child L2s?

• What are the general cognitive correlates of bilingualism in children? These will not be the same as in the case of late (adult) bilingualism, for obvious reasons But again, it may turn out that they do not differ qualitatively Is thought affected in any way by knowing more than one language?

• In L2 learning, especially for academic purposes, what are the effects of L1 edge, and how is conceptual knowledge available to bilingual learners?

These are the topics that the next nine chapters will address

Educators, who in the past have often viewed any language spoken in the room other than the offi cial one as a deviation from the ideal, are today more interested in the opportunities that bilingualism and early L2 learning offer, both

class-• for children who are monolingual speakers of the national language or offi cial language of instruction, and

to plague bilingual and second language education, despite clear policy guidelines

in most school systems, is the disproportionate assignment of L2 learners to special education programs (i.e., assigning children who do not suffer from an intrinsic language disability) Although they may overlap, the school populations served by speech and language therapy and by L2 instruction should be considered different Less egregious “ dis-ability ” approaches to child bilingualism, but unfortunately just

as widespread, include assessment of learning outcomes exclusively through the L2, and uninformed and informal characterizations of “ semilingualism, ” when the cor-responding monolingual condition normally would apply only in cases of true language impairment The identifi cation of such linguistic defi cits requires a formal clinical diagnosis based on appropriate assessment

Even though few educators today would consider young L2 learners as suffering from a kind of impairment, in practice most pedagogical approaches fail to take into account the existing linguistic knowledge of bilingual children Curriculum models and teaching methods are often based on the idea that L2 learners ’ L1 knowledge is either an irrelevant factor or an obstacle to achieving advanced levels

of language profi ciency Indeed, much early research on child bilingualism concluded that early exposure to two languages counts as a potential risk factor with regard

to normal linguistic and cognitive development (for a historical overview, see C Baker 2006)

While few language development specialists today would warn bilingual parents and second language teachers against promoting bilingualism in children, might

there be under some circumstances a secondary or transitory “ cost ” to processing two languages that in turn could affect learning and performance in school-related tasks? D K Oller and Eilers (2002) ask this interesting question Of far-reaching theoretical importance, it also comes up often in the real world when teachers are asked to counsel parents whose children are experiencing delays in language devel-opment and initial literacy learning This is in fact one of the language-learning problems about which we want to keep an open mind

A good place to start this discussion would be around the complex issue of eracy learning, complex enough in just one language It quickly becomes diffi cult

lit-to sort out the interacting faclit-tors when we consider all the possibilities: literacy in the child ’ s primary language, in both primary languages if bilingualism is simultane-ous, and in the second if bilingualism is sequential, to mention just some of the broad categories Since this is the aspect of bilingual profi ciency with which this book will be most preoccupied, another major theme will be that of children ’ s development of language awareness What is the connection between this kind of awareness, on the one hand, and bilingualism, L2 learning, and literacy, on the other? The problems of bilingual literacy will be introduced in the next two chap-ters The setting that we want to consider in particular is the minority language community where children learn to read and write in their L2

1.1 Bilingual Profi ciency and Bilingual Competence

The title of this book refers to both “ competence ” and “ profi ciency ” The suggestion

is that there is an important distinction here At fi rst, this may seem to be simply a matter of sorting out terms that in common usage are interchangeable A “ compe-tent teacher ” usually refers to someone who is skillful and able and demonstrates

a high level of performance But this is not the way that competence is understood

by most linguists

In this book, “ profi ciency ” will refer to aspects of skill and performance, mous with “ ability ” A language profi ciency test measures performance on a given set of skills Bilingual children might demonstrate degrees of profi ciency in more than one language Moreover, in discussing profi ciency we usually need to be spe-cifi c — that is, to talk about skill or ability in using language for a specifi c purpose: reading, writing, listening, or speaking in L1 or L2

Profi ciency or ability, then, is skill in performance, adeptness in using language

in comprehension or expression The idea of a “ specifi c purpose, ” or any purpose for that matter, also implies that the use to which language is put is meaningful or potentially meaningful in some way For bilingual children, it is the ability to use one or the other language that they know, or even both together, for some meaning-ful purpose The studies reported on here from the Mexican research project have

in fact taken one particular set of purposes as a major point of reference: those linked to school achievement, literacy, and other academic uses of language “ Competence ” will refer to something different: knowledge For a user of a language to be able to understand a question or respond with a coherent answer,

he or she must possess linguistic knowledge (among other kinds of knowledge) In this book, this knowledge will be viewed as being specifi c to a particular language,

or particular languages (Note that this aspect of the defi nition is different from that

of many linguists.) Competence, then, is about underlying cognitive structures that store knowledge The intuition that knowledge is not the same as ability comes from the frequently observed inability to put knowledge to use The most dramatic examples are of people who have lost some aspect of language profi ciency but who demonstrably have not lost at least some underlying component of the knowledge needed for the ability Under this category, the most interesting cases are those in which impaired language ability is recovered within such a short time that we would not want to say that the relevant knowledge structures were acquired or learned a second time around In other words, competence is based on mental representations that have a “ content ”

Considering other examples of language breakdown, a patient may suffer from

a condition in which normal speech production is impaired, but in which, in a more controlled setting, he or she can perform perfectly on judgments of grammaticality

In an aphasia that affects only expression or comprehension, but not both (some aphasias do affect both), or a dyslexia that affects reading but not auditory com-prehension, the most likely explanation for the differing performance is that the underlying knowledge of grammar, in part or as a whole, has been spared (Obler and Gjerlow 1999, 144) This also assumes that the same basic grammatical com-petence underlies both production and comprehension (put to use, to be sure, by different processing mechanisms) Otherwise, we would be forced to assign a dif-ferent kind of grammatical knowledge for every kind of language use

To recap: profi ciency = ability, competence = knowledge Note that according to the way the distinction between competence and profi ciency is presented here, lan-guage breakdown could conceivably be traced to one, to the other, or to both in varying proportions

In practice, the distinction between competence and profi ciency is much more complex — a topic of ongoing debate among professional linguists and psychologists, even within the theoretical approaches that accept the distinction in the fi rst place The problem is not new, either A particularly illuminating early attempt to frame the broader question we owe to Plato ’ s allegory of the cave from part III of the

Republic : how do we attain progressively better understanding of the world through

experience? A more modern reading of the allegory might ask us to consider a specifi c problem In regard to how we experience the world around us, there appears

to be an interesting relationship between perception and how information is stored

in memory The input to our senses is passed to a fi rst line of processing mechanisms, eventually fi nding its way to internal mental structures where the essential properties

of this information are stored We could think of these as “ more internal ” than the “ outer layer ” of input/output processing mechanisms For example, visual input is received from all sorts of circular objects in an infi nite variety of presentations and

in all degrees of degradation, as in their orientation, projecting elliptical images From this inconstancy the mind constructs constant representations, as if it pos-sessed ideal models for different categories of phenomena In Plato ’ s cave, projec-tions of constant forms (themselves models of real objects), now deformed and impoverished, are all that the captive perceivers have access to Plato speculates on the possibility of one prisoner fi rst ascending out of this den of shadows and study-ing the invariant forms themselves, then climbing out of the darkness to study higher sources of knowledge, in the manner of philosophers and psycholinguists Plato then asks us to consider a pedagogical implication:

Then if he called to mind his fellow prisoners and what passed for wisdom in his former dwelling-place, he would surely think himself happy in the change and be sorry for them They may have had a practice of honouring and commending one another, with prizes for the man who had the keenest eye for the passing shadows and the best memory for the order

in which they followed or accompanied one another Would our released prisoner be likely to covet those prizes or to envy the men exalted to honour and power in the Cave? (Plato 1941, 230)

If this is true, then, we must conclude that education is not what it is said to be by some, who profess to put knowledge into a soul which does not possess it (p 232)

How, from impoverished input, do both our perceptions and our concepts come

to be as rich and complex as they are? In other words, how is it that knowledge is underdetermined by experience? There appear to be two ways, at least, of making the connection to our distinction between profi ciency and competence

First, previous knowledge enriches the input processed through the senses While for native speakers of a language, negotiating a transaction over a bad telephone connection may present no insurmountable diffi culty, the L2 learner may require a face-to-face meeting The L2 learner ’ s knowledge is insuffi cient

to upgrade the input Newmeyer (2008, 119 – 122) gives a different kind of example

of the same general idea: how sentence fragments are understood Even though

in actual speech we may use truncated forms, underlying mental representations embody all the relevant principles and help us make sense of “ incomplete ” phrases Possible answers to the question “ Who does Betty want to wash? ” could be any of these:

“ Herself ”

“ Her ” (not Betty)

“ Him ”

“ Me ”

But these are not possible (coherent) answers:

“ Myself ”

“ Her ” (referring to Betty)

The well-formed fragment fi ts into a pattern that aligns with the form and meaning

of a full sentence:

“ Betty wants to wash herself ”

“ Betty wants to wash her ” (i.e., someone else)

“ Betty wants to wash him ”

“ Betty wants to wash me ”

Fragments aligning with the following sentences are not well-formed (in the sense

of being either ungrammatical or incoherent) or cannot be assigned the intended meaning:

“ Betty wants to wash * myself ”

“ Betty wants to wash * her ” (referring to Betty)

A related practical problem in language teaching is the common misconception among inexperienced second language teachers that “ sentence fragments ” are ungrammatical Apparently neglecting to refl ect on their own, generally grammati-cal, use of “ incomplete sentences, ” native-speaking instructors sometimes give learn-ers misleading advice They often take L2 students to task for pragmatically appropriate truncated forms, instructing them to “ speak in full sentences ” Of course, practice in formulating complete sentences, in the appropriate language use context, is of undeniable pedagogical value for improving academic writing But if corrective feedback is to be effective, teachers should try not to confuse well-formed fragments and true errors

Second, the competence-performance (profi ciency) distinction comes up when an inability to process input or produce well-formed utterances may not necessarily reveal the full contours of a listener ’ s or speaker ’ s competence An interesting inverse relationship observed among bilinguals (perhaps unforeseen by Plato) highlights the distinction in this regard Skill in managing information (a hyperability perfected

by some L2 learners) partially compensates for incomplete knowledge of grammar Literacy tasks in L2 lend themselves ideally to the application of such a strategy This L2 learner “ trick ” often creates the impression that competence is more advanced than in fact it is (Lebrun 2002) In this example, competence doesn ’ t

advance, yet; rather, performance is boosted by more highly developed processing skills

Many years after Plato, Descartes again posed the problem of how knowledge and experience are related: specifi cally, suggesting that the former is not likely to

be simply a matter of induction, of forming generalizations and concepts from examples provided by sensory input alone:

But as for the essences we know clearly and distinctly, such as the essence of a triangle or any other geometric fi gure, I can easily make you admit that the ideas of them which we have are not taken from particular instances [When] in our childhood we fi rst happened

to see a triangular fi gure drawn on paper, it cannot have been this fi gure that showed us how

we should conceive of the true triangle studied by geometers, since the true triangle is contained in the fi gure only in the way in which a statue of Mercury is contained in a rough block of wood Thus we could not recognize the geometrical triangle from the diagram on the paper unless our mind already possessed the idea of it from some other source (Descartes 1984, 261 – 262)

Modern-day psychologists, some of them following Plato ’ s and Descartes ’ s lead, have suggested that young children begin to categorize objects by developing an understanding, on some level, about objects ’ more fundamental properties, beyond their outward appearance For Bloom (2001, 1102), the research on early concept formation and language development points to a “ rationalist account ” of word learning; that “ children ’ s categorization, and their use of words, is governed by an essentialist conceptual system ” In childhood, surely, there are many examples of how we seem to “ know ” much more than we should, given what experience pro-vides If correct, this would be true twice over in bilingualism Consider the follow-ing mixed utterances from a 2;5-year-old Spanish-English bilingual (Elizabeth in appendix 3):

Oo ta Papi ’ s coins?

Titas de agua in the baby tree

One arriba , one abajo , one abajo , one arriba , one here

Baby talk for:

[ Where is Daddy ’ s coins?]

[ Little drops of water in the baby tree.]

[One up , one down , one down , one up , one here.]

Researchers studying child bilingual development have pointed to similar systematic patterns of switching and borrowing from numerous language pairs as evidence for the early separation of the grammatical systems For example, phrases tend to be kept intact, in the same language; and word order patterns generally match up at switch points This differentiation of language subsystems, internal to the Faculty

of Language, 1 appears to complete its course far in advance of any declarative knowledge or conscious awareness of the differences between the languages (Genesee, Paradis, and Crago 2004) Switching appears to be rule-governed, an expected consequence of early separation of the two grammatical subsystems For example, balanced Spanish-English bilingual 2-year-olds would tend not to produce patterns

like “ Where ta Papi ’ s monedas ” or “ Titas de water in el beb é tree ” (although such

utterances that contravene the word order of one or the other language are by no means impossible, especially if one language is dominant) This kind of internally regulated interaction between language subsystems is further evidence that young children ’ s developing mental grammar is more elaborate and structured than their utterances suggest (Newmeyer 2008) Research has shown that this specialized linguistic knowledge emerges spontaneously despite the fact that the child receives language input from “ competing ” sets of examples, input that one might suspect is

at least potentially confusing In addition, mixing tends to be systematic even in the absence of well-formed codeswitching models, or any model of language mixing if parents do not codeswitch in the child ’ s presence, as attested in Elizabeth ’ s case (see appendix 3)

In different ways, then, knowledge enriches and upgrades the input we receive from the outside world However, just as the higher domains of Plato ’ s cave were not hermetically closed off from the shadow world, so abstract categories, geometrical concepts, and existing linguistic knowledge should work closely with incoming information Indeed, the distinction between knowledge structures and processing modules may not be as clear-cut as it is portrayed here However,

it makes a good starting place, one we will pick up on in the chapters to come

In sum, the way the relationship between knowledge and ability is being

conceived here is that there is a relationship, in fact a close one From the point

of view of understanding a particular category of related abilities (e.g., related), it would be odd to start with the idea of a sharp disconnect between ability and knowledge Taking an example of a specifi c academic language ability among the many that children learn in school, reading comprehension, one proposal would be that competence is an integral component of profi ciency But

school-it is apparent that competence school-itself should be disaggregated In reading sion, a number of interacting competencies must come together, be accessed, and

comprehen-be coordinated, very quickly Processing mechanisms and interfaces of different kinds must effect this coordination Some may be specifi c to certain knowledge structures, and others may be nonspecifi c processors of the type called “ domain-general ”

The same type of relationship should apply to an ability that is acquired without instruction, such as profi ciency in conversation In the case of a bilingual, this would

include the ability to sustain a coherent face-to-face conversation using two languages Profi ciency in this kind of communication also requires the concurrence

of various competencies (of both linguistic and nonlinguistic knowledge), nated with the help of the information-processing components of the skills in ques-tion In this sense, perhaps, a more precise title for this collection of studies would

coordi-have been Bilingual Competencies and Profi ciencies , plural Or, from another point

of view, one could say that in the study of any one bilingual profi ciency, considered

in isolation from others, the former is singular, the latter plural: “ competencies ” But none of this makes for a good book title

On a related note, years of research on the effects of social, cultural, and nomic factors on language development have made clear that “ language ” is in fact

eco-a broeco-ad umbrelleco-a term In regeco-ard to some eco-aspects of “ leco-angueco-age, ” the effects of these external factors have been diffi cult to observe or measure in a reliable way In fact, their infl uence on certain core components of grammatical knowledge can be shown

to be indirect and even secondary But in regard to other aspects (including even some components of what we call “ grammar ” ), the effects are more demonstrably broad and thorough-going If this is true, it should be cause for serious refl ection about the possible componential nature of language ability and language knowl-edge How does the research panorama on these very hard questions change when

we consider two languages instead of just one?

A full discussion of these questions goes far beyond the scope of the coming chapters The idea behind this overview of the problem is to lay out the fi rst of several tentative proposals: that while knowledge/competence should not be equated with ability/profi ciency, it would be unwise to draw the distinction out to the point where the two are completely divorced When competence is actually put to use, typically multiple competencies come together to serve a given ability A necessary implication of this relationship (and for the study of language development, a happy one) is that evidence from performance always reveals something about competence,

to varying degrees of indirectness For example, fi ndings from psycholinguistic research that call into question a theory of linguistic competence should be taken into account seriously The evidence may turn out to be too indirect; in the experi-ment too many cognitive general factors could not be controlled for, maybe But

fi ndings from rigorously conducted studies of language use that plausibly bear on questions of language knowledge are important for theorists even when results are inconclusive

So far, three working hypotheses have been proposed as a way of setting the stage for the discussions to come:

• the early separation of the language subsystems in bilingual children is an example

of how competence does not depend completely on experience and how it in turn may be separable into components

In the coming chapters, we will weigh some of the evidence that might support these proposals In doing so, we will consider whether certain strong claims associated with some of them might be too strong If empirical fi ndings compel us to retreat toward weaker hypotheses, this cannot be a bad thing, for one because it allows us

to pursue more seriously an interdisciplinary approach to the problems of bilingualism

1.2 Knowledge That Outstrips Experience

A twenty-fi rst-century controversy that Plato and Descartes anticipated in the

pas-sages quoted earlier turns on different approaches to what is called the stimulus problem To take one example: the knowledge acquired by all normally

poverty-of-developing children appears to far outstrip the evidence available to them from the environment alone (the so-called primary linguistic data ) In other words, the

knowledge attained seems to be underdetermined by the totality of positive ples encountered in the primary linguistic data The number of alternative sets of grammatical principles that are consistent with the regularities that an “ empiricist learner ” would notice are just too numerous Given how quickly children show evidence of knowledge of one particular set of patterns, there is good reason to start with the hypothesis that some kind of acquisition mechanism helps them zero in on what they have to know

For children who receive input in two languages, the language subsystems are differentiated at an earlier age This suggests the operation of language acquisition mechanisms that not only are specialized for processing linguistic input but also can accommodate this input from two different grammars at the same time Even if the separation is not complete at a given stage of development, the degree to which young children begin to form autonomous linguistic subsystems along the way is diffi cult to account for with a model based on the application of purely inductive learning strategies We could even venture to say that for dual-language acquisition the “ stimulus ” suffers from a different kind of “ poverty, ” one that is potentially more diffi cult to surmount: too much data from which to construct two separate grammars Nevertheless, young bilinguals hit on the right hypotheses separately for each language subsystem Thus, the poverty-of-stimulus problem has led to the postulation of acquisition processors that are domain-specifi c (S Anderson and Lightfoot 2002; Laurence and Margolis 2001) We will encounter this idea again

in chapters 5 and 6, specifi cally with regard to how it might apply to bilingualism

and L2 learning In the meantime, a question to keep in mind is this: might there

be some domains of language and language-related knowledge to which the of-stimulus problem does not apply?

This tradition of theorizing about how language is acquired poses an alternative

to the idea that a single generic associative learning process can account for the richness and complexity of human knowledge of language It would be incorrect to deny that associative learning and inductive strategies play a role in language devel-opment Indeed, such a denial would be unsustainable in the face of self-evident facts regarding language ability in a number of domains Rather, the argument is that there needs to be a certain minimum level of built-in cognitive machinery for young children to be able to process linguistic input for acquisition as rapidly and effectively as they do Linguistic data have to be processed in such a way that the result is not only linguistic competence, but multilingual competence — all this in cognitively immature learners who still have trouble fi guring out things that appear

to be much less challenging

The modern-day version of the alternative to associationism is described by Pinker (2002) as the “ computational theory of the mind ” Sense organs transduce physical energy into patterns of activity and initial confi guration: information Intermediate processors structure these patterns further In turn, thinking and rea-soning can also be thought of in terms of computation and transformation These kinds of cognitive procedure require the concurrence of different kinds of informa-tion-processing system, each internally structured for specialized mental operations that transform information Basic capacities that underlie the construction of lin-guistic knowledge include:

1.3.1 Analysis of Language Abilities

The profi ciency-competence distinction and the poverty-of-stimulus problem lead

to the second recurring theme of the book: the concept of modularity — how an ability, for example, can be analyzed componentially “ Components ” of language ability can be understood in two ways We could refer to components simply as a matter of terminological convenience, to make it easier to talk about grammar (the “ parts of speech ” ) or to analyze it as an object of formal description An example

would be to describe the rules of a certain dialect for the purpose of prescribing “ proper usage ” The second way to understand “ components ” is that they are good for more than just chapter titles in books about language They are actual cognitive structures, materially: subsystems of cognition that are psychologically real — that

is, correspond to neural networks in the brain Modular approaches to the study

of bilingualism all share the second notion of “ component ”

Bilingual development in children should offer greater possibilities for testing different hypotheses related to the claim that language profi ciency and language competence are modular in some way The emergence of two distinct linguistic subsystems in child bilingualism (what we call “ two languages ” — e.g., French and Spanish) presents the fi rst set of research questions related to the concept of modu-larity If each language subsystem, by itself, can be further analyzed into component subsystems (i.e., if the different domains of linguistic knowledge and processing capability can be broken down into components), which of these subsystems in each language develops independently, and which might have components shared in common? Or are all aspects of language in bilinguals represented in a completely integrated and undifferentiated manner? Or are all components and subcomponents language-specifi c, with no shared domains? In other words, how and to what degree might the two language subsystems of a bilingual be separate, if in fact they are? Is this separation manifested, (micro)anatomically, in actual neurolinguistic differen-tiations; and if so, how is it achieved developmentally?

A few informal examples might help to make the notion of modularity more concrete First, we could compare the abilities of two bilingual children in a hypo-thetical 3rd-grade classroom in an isolated mountainous region in South America Both speak the indigenous language of their community, and both started to learn Spanish in 1st grade In regard to mastery of the basic grammar of their L1 and L2, they are indistinguishable However, Asenci ó n (A) has been steeped in the oral tradi-tion of his community through daily exposure to it in his family from an early age Bonifacio (B) has had virtually no contact with traditional narratives, ceremonial discourse, religious sermons, or popular poetry and songs In and out of school, A demonstrates exceptionally high levels of mastery in tasks related to academic-type discourse ability For example, in Spanish (his L2) he can produce, coherently and skillfully, a complex narrative with multiple characters, attributing internal psycho-logical states to them, and he can generate embedded story lines But at the “ sentence level, ” his knowledge of Spanish grammar is rudimentary Meanwhile, B exhibits exactly the converse profi le in his L1: fl awless grammar and rudimentary narrative ability University students sometimes comment on this kind of “ double dissocia-tion ” in comparing a favorite professor, a nonnative speaker of the language of instruction, with a native-speaking professor whose lectures are hard to follow (not because of faulty grammar or diffi cult pronunciation)

In another domain of language, we could compare our eloquent foreign college professor with a native speaker of the professor ’ s L2 — say, one who has never received formal education Since the professor studied the L2 throughout her sec-ondary school, university, and postgraduate years, she has built up a sizable vocab-ulary, more extensive in fact than is typical of native speakers of the language However, while her absolute number of L2 lexical entries is large, at the word level she experiences persistent diffi culties with infl ectional affi xes and other aspects

of grammar related to word formation The non-university-educated speaker of the language displays error-free mastery of L1 infl ectional morphology, as we would expect, but for reasons that are also easy to explain he has access to a less extensive vocabulary But independent of the absolute number of entries in his L1, and in contrast to the L2 learner, each entry in the native speaker ’ s lexicon is more complete and well-formed (Van de Craats (2003) analyzes what “ lexical knowl-edge ” consists of, in L1 and L2, and Lebrun (2002) explains how different aspects

of the lexicon develop (and break down) unevenly.) This dissociation is also related

to the distinction between implicit and explicit knowledge systems We will sider a similar illustrative anecdote in chapter 5 after we see how this kind of internal bilingual differentiation might be represented graphically (in fi gures 5.2 – 5.4)

As we discuss literacy learning, the idea that an ability can be analyzed nentially will be useful in a number of ways Coltheart (2006), for example, empha-sizes the need to study the “ individual components of the reading system, ” especially

compo-in trycompo-ing to understand the causes of readcompo-ing disability:

When children are learning to read, they are acquiring a mental information-processing system that they previously lacked, and that has some particular functional architecture or other: the system will have as components distinct modules, each responsible for one of the different information-processing jobs that have to be carried out for an act of reading to be successfully accomplished (p 124)

Beginning in this introductory chapter and continuing in chapters 3, 4, 8, and 9,

I will present fi ndings from an ongoing research project on bilingual literacy opment that colleagues and I carried out Comparisons between measures of chil-dren ’ s performance on school-related literacy tasks, in both languages that they understand, called our attention to the concept of modularity The question that came up very quickly in this work was how a truly analytical approach can be applied to the study of literacy development in children ’ s L1 and L2 What are the components and subsystems that make up this one aspect of bilingual profi ciency?

devel-In an attempt to get a grip on this question, we made this provisional assumption: understanding the complex interactions among these component domains calls for

a working model that at least tentatively assumes that some of them enjoy some degree of autonomy

How this working hypothesis fares will depend on future investigations very ferent in kind from the studies we were able to conduct in school settings, far from the controlled conditions of experimental research In the end, the hypothesis may turn out to be wrong But the idea was to start with a reasonably plausible frame-work and then take it as far as it will allow us to go, until it breaks down and leads

dif-to different questions Basic questions that still need dif-to be formulated properly include these:

• How are these kinds of interaction constrained?

1.3.2 The Idea of Components Applied to the Problem of Language Learning

in School

The starting point for discussing a working model of this sort was Cummins ’ s framework for how bilingual profi ciency in the academic realm should be under-

stood In its current version (Cummins 2000, 191), the Common Underlying

Pro-fi ciency/Central Processing System ( CUP/CPS — henceforth CUP ) model suggests an

interdependence between L1 and L2 “ Interdependence ” means that some nents are more language-specifi c (have separate representations for each language) and that others belong to domains that are shared or “ overlapping ” This distribu-tion between language-specifi c domains and shared domains is consistent with research on the cognitive architecture of bilingualism, in particular, with many of the seemingly contradictory fi ndings For example, in bilinguals, some aspects of language ability appear to be based on a common, undifferentiated store of knowledge, while other aspects appear to depend on knowledge structures that are specifi c to L1 or L2 (for reviews of the literature, see Heredia and Brown 2004;

compo-M Paradis 2004)

Chapter 3 will show how the CUP model accounts for key aspects of dence between L1 and L2 in the performance of language tasks related to literacy

interdepen-In an attempt to portray the language-specifi c and shared domains more precisely,

a fi rst approximation of a modifi ed version of Cummins ’ s model will be proposed Chapters 4 and 5 will then review current research, leading to the conclusion that the modifi ed version is inadequate in important respects The guiding idea in this evolution of proposed models is the need to seriously examine the distinction between language-specifi c and shared domains, and how and to what extent they are interconnected The concepts of modularity, poverty of stimulus, and the com-petence-performance distinction are most commonly associated with Universal

Grammar and generativist approaches to the study of language But none of these concepts draw a sharp line that should lead us to question the applicability of research fi ndings from the various functionalist approaches, in particular in the applied fi elds Keeping this in mind is especially important given our primary interest

in bilingual profi ciency, aspects of performance, and L1-L2 language use From this starting point, the possibilities for exploring new and unexpected intersections of consensus might be more promising than they fi rst appear It is no secret, even among its most fervent proponents, that Universal Grammar – oriented research on bilingualism and L2 learning has suffered a growing isolation within the fi eld Numerous investigators fi nd nothing relevant in either the specifi c theoretical models

or the general concepts outlined so far in this introduction Starting with chapter

3, as we begin to consider what kind of evidence might falsify the claims of larity, poverty of stimulus, and the competence-performance distinction, there will

modu-be occasion to refl ect on why these notions have fallen into disfavor As noted earlier, part of the reason might lie in the tendency to insulate some of the guiding constructs

of the Universal Grammar approach from potentially disconfi rming fi ndings in the broader realm of the cognitive and social sciences

In fact, functionalist-oriented (emergentist, connectionist) researchers have tributed in two important ways to better understanding the issues in the study of bilingual profi ciency: (1) Some lines of criticism have undermined versions of modu-larity, the competence-performance distinction, and parameter-setting models of acquisition that can be characterized as too narrow or too radical Most functional-ists might wish to point out that their intention has been to question the assumptions

con-of generative grammar in a more fundamental way But for our purposes, these critical observations have helped clarify basic constructs related to dual-language ability, mainly by suggesting that we might want to step back from hypotheses that are too strong E Bates, E Clark, H Diessel, A Goldberg, A Karmiloff-Smith, and

M Tomasello belong to this fi rst group (2) The actual empirical fi ndings and ses from connectionist and functionalist research bearing on important applied questions have coincided — unexpectedly, I might add — with Universal Grammar – oriented researchers ’ conclusions about the same questions Specifi cally, these coin-cidences have appeared in the areas of L2 learning and literacy learning Explanations may not be converging in the strict sense, but the concurrences raise important questions For one, might there be a certain common ground, conceptually (not just

analy-in regard to practical applications), that accounts for these poanaly-ints of contact? Among this group, B MacWhinney, M Schleppegrell, and M Seidenberg come to mind Readers will recognize other authors mentioned in the coming chapters who fall into one of these two categories This theme deserves an extensive discussion

of its own Introducing it in this book, albeit in a very summarized way, I hope encourages some much-needed refl ection on future directions in bilingual research

1.4 A Study of Indigenous-Language Bilingualism in Mexico

1.4.1 Evaluation of Academic Language Profi ciency and Literacy Learning

What makes child bilingualism especially informative in studying the components

of language profi ciency is the wide variation in types of contact situation between L1 and L2, and all subsequent L2s The extensive variability in social context as it affects the use of two or more languages can push certain language-learning circum-stances to the limit In this regard, we can take the ideal situation of social equilib-rium between languages as a hypothetical standard against which to compare varying degrees of disequilibrium

In North America, for example, there is generally a lesser degree of imbalance

in French literacy learning in Anglophone Canada than in Spanish literacy learning

in most places in the United States And the relation between Spanish and indigenous languages in Latin America represents a qualitatively different order of imbalance Even among the indigenous languages, a clearly discernable hierarchy vis- à -vis Spanish ranks major, offi cially recognized, and historically valued autochthonous languages above those with fewer speakers, diminished ethnolinguistic vitality, and lower prestige In an interview done many years ago, an elderly gentleman from the town in which our study was carried out offered his view of the local hierarchy Beginning with a comparison between his mother tongue, Nahuatl, and another indigenous language spoken in a nearby town, Nh ä h ñ u, he remarked that the former was much more useful in the immediate vicinity of the rural towns of the region More importantly, it was a language of higher culture, of a former grand empire, with a literature Nh ä h ñ u was more like the “ dialect ” of a “ tribe, ” one that hap-pened to be also in an advanced stage of extinction in his state of Mexico Spanish stood above all as the national language, useful especially for schooling; English, now immensely popular among young people, for communication globally; and Latin for the highest spheres on Sunday mornings Similar social disequilibrium in regard to different realms of language use is evident in virtually all multilingual societies These relations allow the possibility of studying, at the level of the indi-vidual language learner,

Puebla ( fi gure 1.1 ) Not far from a major industrial corridor that runs in a number

of directions connecting other major urban areas, and home to one of Latin ica ’ s largest auto assembly plants, this community is distinguished by a high level

Amer-of bilingualism This contact situation appears to be related to the coincidence Amer-of extensive daily access to the national language (especially on television) with a not yet fully explicable maintenance of the indigenous language The specifi c locality of San Isidro and San Miguel where the study was undertaken is the most linguistically conservative of all indigenous-language-speaking communities in this region For example, systematic assessment of language dominance among a sample of older children (in 4th and 6th grades) revealed a surprising profi le: no monolinguals or incipient L2 learners of either Spanish or Nahuatl Balanced bilingual conversational ability was the norm, with no evidence of nonnative grammatical competence among the older children except for one Spanish-dominant 4th grader who, never-theless, could be described as an intermediate speaker of Nahuatl Extensive assess-ment in both languages, confi rmed by child and parent interviews, provided the

Figure 1.1

The Puebla-Tlaxcala valley

following profi le: the few children who enter 1st grade with nonnative or incipient knowledge of Nahuatl (typically “ passive bilinguals ” ) acquire at least intermediate conversational profi ciency by 4th or 5th grade Similarly, Nahuatl-dominant 1st graders (also a minority, but more numerous than the Spanish-dominant speakers) rapidly approach levels of balanced bilingualism

Three sets of literacy-related language assessments were administered in one of the public elementary schools, designated offi cially as “ bilingual, ” 2 that draws stu-dents from both of the above-mentioned adjacent towns that form part of the community:

1 a broad descriptive survey of language and literacy performance, in both guages, in a selected cohort of 2nd, 4th, and 6th graders;

2 a follow-up assessment to confi rm observed tendencies in an unselected group

of 3rd and 5th graders;

3 among the original cohort, a more narrowly focused evaluation of one aspect of profi ciency in reading and writing: self-correction

This chapter will set the stage with a summary of the fi rst part of the study:

a description of tendencies across grades 2, 4, and 6, in conversational discourse, oral narrative (coherence and completeness of story construction), written narrative (same criteria), and reading in Spanish and Nahuatl (oral miscue analysis, retelling, and a cloze test) Participants were selected by teachers with the purpose of examining normal grade-level achievement Therefore, performance

on academic language tasks was expected to show upward tendencies from 2nd

to 6th grade in Spanish, the medium of instruction, as in fact it did These cies were statistically signifi cant between 2nd and 6th grades in all three literacy-related areas (not in conversational discourse, which should not be expected to show variation after early childhood) Conversational abilities in both languages were sampled as part of the assessment of language dominance, and predictably, no grade-level trends and no appreciable differences between Spanish and Nahuatl appeared

In school, the indigenous language received positive recognition, was broadly tolerated and respected, and was the object of a demonstrably forthright symbolic valorization, in line with offi cial national educational policy for bilingual programs Perhaps because most children entering 1st grade in this community command at least passive comprehension abilities in Spanish, and because of other pressures and constraints internal to the community, literacy instruction is almost exclusively carried out in the national language Exceptions include occasional writing contests

in Nahuatl and introduction to the Nahuatl alphabet

To reiterate, especially given the teachers ’ selection criteria, literacy testing

in Spanish, the language of instruction, was expected to show typical grade-level

progressions In comparison, two equally plausible outcomes in response to the parallel Nahuatl assessments could be predicted:

1 Performance in the sociolinguistically disfavored indigenous language, virtually excluded from day-to-day literacy instruction and practice, lacking any compensa-tory reading material in the community or at home, and with minimal presence in “ environmental print, ” would show weak, nonsignifi cant advances across the grade levels, or none at all

2 Statistically signifi cant ascending curves would be found, commensurate with the achievement curves in Spanish, despite the sociolinguistic imbalance

Results revealed, for all three academic language measures, without exception, tendencies consistent with hypothesis #2 (see fi gure 1.2 ) Between 2nd and 6th grade, none of the curves for Nahuatl patterned as hypothesis #1 would predict The negative effects of minimal language-specifi c literacy instruction and practice

in the socially disfavored language appear to be much weaker than might have been anticipated In other words, the overall absence of the indigenous language from literacy-related learning activities in school might have (or should have) had a greater depressing effect on performance on literacy tasks when these were evaluated

2nd grade

4th grade

6th grade

2nd grade

4th grade

6th grade

Figure 1.2

Performance of bilingual students on academic tasks in Spanish and Nahuatl

in the indigenous language However, a different result would have been expected

if students were beginning L2 learners of Spanish or beginning L2 learners of Nahuatl, or one-time native speakers who had experienced early attrition and were now strongly Spanish-dominant Note also that while Nahuatl performance on school-related tasks “ parallels ” Spanish performance in the sense that signifi cant differences resulted in all three grades, impressionistically Nahuatl appears to advance less robustly than Spanish In every case, the “ scissors ” open toward 6th grade (N Francis and Nieto Andrade 1996) Speculations regarding this effect will

be taken up in chapter 3 Summing up for now: In some respects, the sociolinguistic situation of these bilingual children affects their performance in noteworthy ways But in certain more important ways, it offers lessons and examples that all bilingual educators and researchers should fi nd applicable to their work In some important ways, their performance is not that exceptional after all

1.4.2 Special Circumstances of Language Learning

In other ways, indigenous languages do provide special opportunities to study guage contact and bilingualism in part because the social inequalities that separate the languages in contact vary widely, sometimes in the extreme Exceptional circum-stances of material economy and access to technology even pose the question of cognitive and linguistic exceptionality With this in mind, the reader is encouraged

lan-to consider the results from a fi eld study of an indigenous-language-speaking munity very different from the one described here: the Pirah ã of the Amazon region (Everett 2005) In fact, the two communities differ radically in almost every way; but this is not of primary concern to us yet Everett ’ s fi ndings should be consulted

com-fi rst because they point to general conclusions that differ fundamentally from the model of language ability that we will consider here In this way, they stand as a pointed challenge Second, in keeping with the idea of evaluating functionalist-oriented research in order to self-critique traditional generative approaches to lan-guage competence, we want to try to account for some of the strong conclusions

of the Pirah ã study One question that Everett raises is, what are the basic, pensable design features of grammar shared by all languages? For example, is recur-sion one of them? Pinker and Jackendoff (2005) present an interesting discussion

indis-of Everett ’ s claims on this point

1.5 Looking Ahead: Overview of the Chapters

There are two threads in the discussion that will follow this introduction, like lel story lines of a narrative One is the report and analysis of fi ndings from a comprehensive evaluation of literacy-related abilities in two languages in the special circumstances of national language – indigenous language bilingualism In some

paral-ways, this thread is indeed like a story, as it opens a window onto the life in school

of a group of bilingual children, profi les their abilities in each language, and samples what they themselves think about the languages they know The other thread, inter-woven among the chapters of the fi rst, is a survey and assessment of the research literature that bears more directly on the theoretical topics introduced earlier: modu-larity, the poverty-of-stimulus problem, and the relationship between competence and profi ciency The fi rst thread should be most useful to bilingual educators and researchers who work in school settings But in the end, the book comes down to

a proposal for looking seriously at the three concepts in the second thread While the research fi ndings reported here do not prove that any one version of modularity

is correct, they do raise the possibility that further research might In any case, proposals about concepts that have yet to attain broad consensus are useful in describing and beginning to understand things The second thread invites research-ers from outside the confi nes of bilingualism and bilingual education to look at new opportunities in the applied disciplines of language learning and language teaching

The discussion in chapter 2 continues the themes of this introduction Researchers interested in cognitive science approaches to bilingualism often lose sight of why anyone else might be interested in the internal workings of mental entities We should not forget about the potential broader applications of work centered on developmental issues involving school-age children, to name just one area of concern

to practitioners For example, if a better understanding of bilingualism and L2 learning eventually could lead to a consensus on more effi cient L2 teaching approaches, this would represent a major practical contribution of linguistic and psychological science This advance would have a cascading effect in a number of areas, including facilitating literacy learning in L2s, providing earlier and more complete access to academic content, and developing other higher-order cognitive abilities associated with schooling Both applied and theoretical researchers would benefi t The advance would open up discussion in the fi eld of educational practice

to new ideas (in short supply in some parts) At the same time, it would provide a real-world corrective in regard to which scientifi c questions are truly important Finally, well-designed studies of language teaching and related macrolevel sociolin-guistic phenomena can contribute to formulating critical alternative hypotheses for psychologists and linguists This chapter ’ s discussion of bilingual education policy worldwide sets the stage for looking at a local example in chapter 3

The Mexican study of academic language ability in Spanish and Nahuatl, reported

in chapter 3, provides some confi rming evidence for Cummins ’ s CUP model Chapter

3 also presents the fi rst attempt at a more explicit version of this model Although the proposed modifi cation will have to be abandoned in chapter 5, it deserves

a moment ’ s consideration, the idea being that the conceptual subcomponents of

language profi ciency, which are not language-bound, need their own domain If something is really “ shared, ” it shouldn ’ t reside with or be the property of either component The options are belonging to neither or to both This problem only gets

a partial solution in chapter 3, but with the door left open for the proposed model ’ s next, more elaborated, version All the versions, by the way, owe their usefulness

in modeling bilingual profi ciency to one idea: the many multidirectional interactions between L1 and L2 are not all of the same kind The metaphor of “ transfer ” has captured the nature of these interactions; at the same time, it has introduced a simplifi cation that sometimes makes things harder to understand

The Mexican study raised more questions than a descriptive investigation of limited scope could even begin to approach Chapters 4 and 5 start to look beyond the confi nes of that project The studies of bilingualism that are reviewed

in these chapters are of an order that our project could never have gathered the resources for, materially or theoretically Hence, our most sincere gratitude goes

to the many authors cited in this book for the meticulous reconstruction, for our benefi t, of their investigations The most suggestive are those that delve into the fascinating realm of exceptional bilingualism, in which atypical conditions

of language development are truly pushed to the limit Chapter 5 concludes with

a presentation of a bilingual version of Jackendoff ’ s (1997) Tripartite Parallel Architecture

Chapter 6 takes up one of the big topics in the study of bilingualism today: the question of critical period effects in L2 learning It is proposed that the debate around nearly every aspect of this question might benefi t from a new lens: results from the emerging fi eld of study into child language attrition Far from merely documenting a phenomenon restricted to small, endangered languages, the study of how language loss unfolds reveals universal properties of linguistic competence Bilingual educators might recognize this phenomenon under the name of “ subtrac-tive bilingualism ” For many years, the fi eld has shied away from a serious examina-tion of this particular outcome of child bilingualism, treating it as something simply

to avoid in its commitment to the “ additive ” alternative One unfortunate side effect has been failure to recognize fully how pervasive, normal, and even typical the subtractive variant actually is in many L2-learning situations Neglecting the scien-tifi c study of L1 attrition in this way has led many practitioners to fl irt with unclear notions such as “ semilingualism ” This happens, for example, when trying to make the (at times exaggerated) case that subtractive bilingualism is always disruptive and damaging to children ’ s development For minority language communities concerned about preserving their linguistic heritage, understanding language attrition should

be important from the point of view of designing an effective language preservation program

Chapter 7 proposes that academic language profi ciency rests upon a foundation formed by two major dimensions of ability: secondary discourse ability and meta-linguistic awareness The chapter focuses on studies of bilingual and L2 literacy and provides a framework for distinguishing one kind of ability from another It also concludes the discussion of the competence-performance distinction with some concrete examples

Chapters 8 and 9 return to the original study of the Nahuatl-speaking community

in Mexico Chapter 8 examines an aspect of literacy learning for which guistic awareness — refl ection on language forms — should fi gure in an important way: self-correction How do bilingual children respond when asked to shift their attention to orthographic, grammatical, and discourse-level patterns in their own writing? How do they respond when asked to do this both in the language in which writing skills are normally practiced in school, and in their “ nonacademic ” lan-guage? Chapter 9 then explores how children shift to a focus on form in reading Chapter 10, aside from giving cause for optimism about future directions in research, questions the usefulness of some popular approaches to bilingualism in school that run counter to the modular perspective Current evidence from various quarters argues rather convincingly that this perspective should at least be consid-ered with an open mind Specifi cally, the critique in chapter 10 centers on strong versions of the “ whole-language ” philosophy This book proposes that the study of bilingualism might advance more productively if it began with an analysis of the components of bilingual ability These components are best thought of as internally structured subsystems, in interactive connections of different kinds Crucially, however, the components and connections cannot do whatever they want; and “ central ” cognitive domains cannot direct the lower-level components any way they want to, either This componential approach should provide a better understanding

metalin-of bilingual children ’ s abilities, more reliable and meaningful assessments metalin-of their abilities in school, and more effective teaching models, above all in literacy and L2 learning

The appendices offer new data to contemplate, data that students of bilingualism are invited to try their own hand at analyzing The data illustrate key concepts discussed in the chapters, along with some preliminary interpretations The appen-dices include (1) procedures for evaluating an aspect of metalinguistic awareness related to children ’ s knowledge of two languages; (2) a summary of fi ndings from

a broad range of measures of developing additive bilingualism, including indices of lexical borrowing, shifts in language dominance, ethnolinguistic loyalty, and obser-vations of language use in conversation; (3) codeswitching data from early simul-taneous bilingualism; and (4) writing samples from Nahuatl-Spanish bilingual students